Auto merge of #75278 - cuviper:indexmap, r=Mark-Simulacrum

Upgrade indexmap and use it more

First this upgrades `indexmap` to 1.5.1, which is now based on `hashbrown::raw::RawTable`. This means it shares a lot of the same performance characteristics for insert, lookup, etc., while keeping items in insertion order.

Then across various rustc crates, this replaces a lot of `Vec`+`HashMap` pairs with a single `IndexMap` or `IndexSet`.

Closes #60608.
r? @eddyb

Author: bors

Cargo.lock

@@ -1382,9 +1382,13 @@ dependencies = [
 
 [[package]]
 name = "indexmap"
-version = "1.0.2"
+version = "1.5.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7e81a7c05f79578dbc15793d8b619db9ba32b4577003ef3af1a91c416798c58d"
+checksum = "86b45e59b16c76b11bf9738fd5d38879d3bd28ad292d7b313608becb17ae2df9"
+dependencies = [
+ "autocfg",
+ "hashbrown",
+]
 
 [[package]]
 name = "installer"

src/librustc_codegen_llvm/coverageinfo/mapgen.rs

@@ -6,7 +6,7 @@ use llvm::coverageinfo::CounterMappingRegion;
 use log::debug;
 use rustc_codegen_ssa::coverageinfo::map::{Counter, CounterExpression, Region};
 use rustc_codegen_ssa::traits::{BaseTypeMethods, ConstMethods};
-use rustc_data_structures::fx::FxHashMap;
+use rustc_data_structures::fx::FxIndexSet;
 use rustc_llvm::RustString;
 
 use std::ffi::CString;
@@ -76,13 +76,12 @@ pub fn finalize<'ll, 'tcx>(cx: &CodegenCx<'ll, 'tcx>) {
 }
 
 struct CoverageMapGenerator {
-    filenames: Vec<CString>,
-    filename_to_index: FxHashMap<CString, u32>,
+    filenames: FxIndexSet<CString>,
 }
 
 impl CoverageMapGenerator {
     fn new() -> Self {
-        Self { filenames: Vec::new(), filename_to_index: FxHashMap::default() }
+        Self { filenames: FxIndexSet::default() }
     }
 
     /// Using the `expressions` and `counter_regions` collected for the current function, generate
@@ -122,16 +121,8 @@ impl CoverageMapGenerator {
                 let c_filename =
                     CString::new(file_name).expect("null error converting filename to C string");
                 debug!("  file_id: {} = '{:?}'", current_file_id, c_filename);
-                let filenames_index = match self.filename_to_index.get(&c_filename) {
-                    Some(index) => *index,
-                    None => {
-                        let index = self.filenames.len() as u32;
-                        self.filenames.push(c_filename.clone());
-                        self.filename_to_index.insert(c_filename.clone(), index);
-                        index
-                    }
-                };
-                virtual_file_mapping.push(filenames_index);
+                let (filenames_index, _) = self.filenames.insert_full(c_filename);
+                virtual_file_mapping.push(filenames_index as u32);
             }
             mapping_regions.push(CounterMappingRegion::code_region(
                 counter,

src/librustc_codegen_llvm/coverageinfo/mod.rs

@@ -97,8 +97,11 @@ impl CoverageInfoBuilderMethods<'tcx> for Builder<'a, 'll, 'tcx> {
     }
 }
 
-pub(crate) fn write_filenames_section_to_buffer(filenames: &Vec<CString>, buffer: &RustString) {
-    let c_str_vec = filenames.iter().map(|cstring| cstring.as_ptr()).collect::<Vec<_>>();
+pub(crate) fn write_filenames_section_to_buffer<'a>(
+    filenames: impl IntoIterator<Item = &'a CString>,
+    buffer: &RustString,
+) {
+    let c_str_vec = filenames.into_iter().map(|cstring| cstring.as_ptr()).collect::<Vec<_>>();
     unsafe {
         llvm::LLVMRustCoverageWriteFilenamesSectionToBuffer(
             c_str_vec.as_ptr(),

src/librustc_data_structures/Cargo.toml

@@ -11,7 +11,7 @@ doctest = false
 
 [dependencies]
 ena = "0.14"
-indexmap = "1"
+indexmap = "1.5.1"
 log = { package = "tracing", version = "0.1" }
 jobserver_crate = { version = "0.1.13", package = "jobserver" }
 lazy_static = "1"

src/librustc_data_structures/transitive_relation.rs

@@ -1,4 +1,4 @@
-use crate::fx::FxHashMap;
+use crate::fx::FxIndexSet;
 use crate::stable_hasher::{HashStable, StableHasher};
 use crate::sync::Lock;
 use rustc_index::bit_set::BitMatrix;
@@ -13,10 +13,7 @@ mod tests;
 #[derive(Clone, Debug)]
 pub struct TransitiveRelation<T: Eq + Hash> {
     // List of elements. This is used to map from a T to a usize.
-    elements: Vec<T>,
-
-    // Maps each element to an index.
-    map: FxHashMap<T, Index>,
+    elements: FxIndexSet<T>,
 
     // List of base edges in the graph. Require to compute transitive
     // closure.
@@ -39,7 +36,6 @@ impl<T: Eq + Hash> Default for TransitiveRelation<T> {
     fn default() -> Self {
         TransitiveRelation {
             elements: Default::default(),
-            map: Default::default(),
             edges: Default::default(),
             closure: Default::default(),
         }
@@ -65,20 +61,16 @@ impl<T: Clone + Debug + Eq + Hash> TransitiveRelation<T> {
     }
 
     fn index(&self, a: &T) -> Option<Index> {
-        self.map.get(a).cloned()
+        self.elements.get_index_of(a).map(Index)
     }
 
     fn add_index(&mut self, a: T) -> Index {
-        let &mut TransitiveRelation { ref mut elements, ref mut closure, ref mut map, .. } = self;
-
-        *map.entry(a.clone()).or_insert_with(|| {
-            elements.push(a);
-
+        let (index, added) = self.elements.insert_full(a);
+        if added {
             // if we changed the dimensions, clear the cache
-            *closure.get_mut() = None;
-
-            Index(elements.len() - 1)
-        })
+            *self.closure.get_mut() = None;
+        }
+        Index(index)
     }
 
     /// Applies the (partial) function to each edge and returns a new
@@ -430,14 +422,11 @@ where
 {
     fn decode<D: Decoder>(d: &mut D) -> Result<Self, D::Error> {
         d.read_struct("TransitiveRelation", 2, |d| {
-            let elements: Vec<T> = d.read_struct_field("elements", 0, |d| Decodable::decode(d))?;
-            let edges = d.read_struct_field("edges", 1, |d| Decodable::decode(d))?;
-            let map = elements
-                .iter()
-                .enumerate()
-                .map(|(index, elem)| (elem.clone(), Index(index)))
-                .collect();
-            Ok(TransitiveRelation { elements, edges, map, closure: Lock::new(None) })
+            Ok(TransitiveRelation {
+                elements: d.read_struct_field("elements", 0, |d| Decodable::decode(d))?,
+                edges: d.read_struct_field("edges", 1, |d| Decodable::decode(d))?,
+                closure: Lock::new(None),
+            })
         })
     }
 }
@@ -452,8 +441,6 @@ where
         let TransitiveRelation {
             ref elements,
             ref edges,
-            // "map" is just a copy of elements vec
-            map: _,
             // "closure" is just a copy of the data above
             closure: _,
         } = *self;

src/librustc_metadata/rmeta/encoder.rs

@@ -4,7 +4,7 @@ use crate::rmeta::*;
 use log::{debug, trace};
 use rustc_ast::ast;
 use rustc_data_structures::fingerprint::Fingerprint;
-use rustc_data_structures::fx::{FxHashMap, FxHashSet};
+use rustc_data_structures::fx::{FxHashMap, FxHashSet, FxIndexSet};
 use rustc_data_structures::stable_hasher::StableHasher;
 use rustc_data_structures::sync::{join, Lrc};
 use rustc_hir as hir;
@@ -48,8 +48,7 @@ struct EncodeContext<'a, 'tcx> {
     type_shorthands: FxHashMap<Ty<'tcx>, usize>,
     predicate_shorthands: FxHashMap<ty::Predicate<'tcx>, usize>,
 
-    interpret_allocs: FxHashMap<interpret::AllocId, usize>,
-    interpret_allocs_inverse: Vec<interpret::AllocId>,
+    interpret_allocs: FxIndexSet<interpret::AllocId>,
 
     // This is used to speed up Span encoding.
     // The `usize` is an index into the `MonotonicVec`
@@ -331,17 +330,7 @@ impl<'a, 'b, 'tcx> SpecializedEncoder<ty::Predicate<'b>> for EncodeContext<'a, '
 
 impl<'a, 'tcx> SpecializedEncoder<interpret::AllocId> for EncodeContext<'a, 'tcx> {
     fn specialized_encode(&mut self, alloc_id: &interpret::AllocId) -> Result<(), Self::Error> {
-        use std::collections::hash_map::Entry;
-        let index = match self.interpret_allocs.entry(*alloc_id) {
-            Entry::Occupied(e) => *e.get(),
-            Entry::Vacant(e) => {
-                let idx = self.interpret_allocs_inverse.len();
-                self.interpret_allocs_inverse.push(*alloc_id);
-                e.insert(idx);
-                idx
-            }
-        };
-
+        let (index, _) = self.interpret_allocs.insert_full(*alloc_id);
         index.encode(self)
     }
 }
@@ -583,15 +572,15 @@ impl<'a, 'tcx> EncodeContext<'a, 'tcx> {
             let mut n = 0;
             trace!("beginning to encode alloc ids");
             loop {
-                let new_n = self.interpret_allocs_inverse.len();
+                let new_n = self.interpret_allocs.len();
                 // if we have found new ids, serialize those, too
                 if n == new_n {
                     // otherwise, abort
                     break;
                 }
                 trace!("encoding {} further alloc ids", new_n - n);
                 for idx in n..new_n {
-                    let id = self.interpret_allocs_inverse[idx];
+                    let id = self.interpret_allocs[idx];
                     let pos = self.position() as u32;
                     interpret_alloc_index.push(pos);
                     interpret::specialized_encode_alloc_id(self, tcx, id).unwrap();
@@ -2019,7 +2008,6 @@ fn encode_metadata_impl(tcx: TyCtxt<'_>) -> EncodedMetadata {
         predicate_shorthands: Default::default(),
         source_file_cache: (source_map_files[0].clone(), 0),
         interpret_allocs: Default::default(),
-        interpret_allocs_inverse: Default::default(),
         required_source_files: Some(GrowableBitSet::with_capacity(source_map_files.len())),
         is_proc_macro: tcx.sess.crate_types().contains(&CrateType::ProcMacro),
         hygiene_ctxt: &hygiene_ctxt,

src/librustc_middle/ty/query/on_disk_cache.rs

@@ -5,7 +5,7 @@ use crate::ty::codec::{self as ty_codec, TyDecoder, TyEncoder};
 use crate::ty::context::TyCtxt;
 use crate::ty::{self, Ty};
 use rustc_data_structures::fingerprint::Fingerprint;
-use rustc_data_structures::fx::FxHashMap;
+use rustc_data_structures::fx::{FxHashMap, FxIndexSet};
 use rustc_data_structures::sync::{HashMapExt, Lock, Lrc, OnceCell};
 use rustc_data_structures::thin_vec::ThinVec;
 use rustc_errors::Diagnostic;
@@ -212,7 +212,6 @@ impl<'sess> OnDiskCache<'sess> {
                 type_shorthands: Default::default(),
                 predicate_shorthands: Default::default(),
                 interpret_allocs: Default::default(),
-                interpret_allocs_inverse: Vec::new(),
                 source_map: CachingSourceMapView::new(tcx.sess.source_map()),
                 file_to_file_index,
                 hygiene_context: &hygiene_encode_context,
@@ -267,15 +266,15 @@ impl<'sess> OnDiskCache<'sess> {
                 let mut interpret_alloc_index = Vec::new();
                 let mut n = 0;
                 loop {
-                    let new_n = encoder.interpret_allocs_inverse.len();
+                    let new_n = encoder.interpret_allocs.len();
                     // If we have found new IDs, serialize those too.
                     if n == new_n {
                         // Otherwise, abort.
                         break;
                     }
                     interpret_alloc_index.reserve(new_n - n);
                     for idx in n..new_n {
-                        let id = encoder.interpret_allocs_inverse[idx];
+                        let id = encoder.interpret_allocs[idx];
                         let pos = encoder.position() as u32;
                         interpret_alloc_index.push(pos);
                         interpret::specialized_encode_alloc_id(&mut encoder, tcx, id)?;
@@ -767,8 +766,7 @@ struct CacheEncoder<'a, 'tcx, E: ty_codec::TyEncoder> {
     encoder: &'a mut E,
     type_shorthands: FxHashMap<Ty<'tcx>, usize>,
     predicate_shorthands: FxHashMap<ty::Predicate<'tcx>, usize>,
-    interpret_allocs: FxHashMap<interpret::AllocId, usize>,
-    interpret_allocs_inverse: Vec<interpret::AllocId>,
+    interpret_allocs: FxIndexSet<interpret::AllocId>,
     source_map: CachingSourceMapView<'tcx>,
     file_to_file_index: FxHashMap<*const SourceFile, SourceFileIndex>,
     hygiene_context: &'a HygieneEncodeContext,
@@ -807,17 +805,7 @@ where
     E: 'a + TyEncoder,
 {
     fn specialized_encode(&mut self, alloc_id: &interpret::AllocId) -> Result<(), Self::Error> {
-        use std::collections::hash_map::Entry;
-        let index = match self.interpret_allocs.entry(*alloc_id) {
-            Entry::Occupied(e) => *e.get(),
-            Entry::Vacant(e) => {
-                let idx = self.interpret_allocs_inverse.len();
-                self.interpret_allocs_inverse.push(*alloc_id);
-                e.insert(idx);
-                idx
-            }
-        };
-
+        let (index, _) = self.interpret_allocs.insert_full(*alloc_id);
         index.encode(self)
     }
 }